PNNL

Abstract

Vegetation greenness characterized by spectral vegetation indices (VIs) is an integrative measure of vegetation leaf abundance, biochemical properties, and composition. Despite greater leaf area index (LAI) in forests, satellite observations of some major VIs are higher over crops (e.g., the Corn Belt in the US) than over the forests (e.g., the Amazon rainforest) during the peak growing season. This contradicting pattern can lead to important misinterpretations in global vegetation studies, thus understanding the underlying drivers and its impacts is essential. Our analysis of multi-sensor satellite observations and radiative transfer model simulations reveals that macroscale shadows cast by complex structures of forests lead to lower greenness measures compared to structurally simple and homogeneous crops. The observed differential response of VIs is inevitable because most Earth Observing satellites do not view the earth in the solar direction and thus view shadows. This shadowing effect is common not only for the reflectance based optical remote sensing observations, but also for the emerging Solar-induced chlorophyll fluorescence. Our finding has significant implications to interpreting VIs in global vegetation changes. We find that regions converting from forests to crops over the Amazon rainforest can show significant decreases in LAI but no changes in the several major VIs due to the shadow impact (i.e. the sudden replacement of forest shadows by more sunlit crops). This divergent pattern is present over global forest and crop regions and could lead to potential misinterpretation of the remote sensing signature over 9.7±3.7 Mkm2 39 area under future land-use change scenarios. This study highlights the importance of taking the shadow impacts into account for correctly interpreting global vegetation changes with remotely sensed VIs.